• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.17 no.2
• /
• pp.113-120
• /
• 2024

The population sparseness index(PSI) is being utilized to describe the functioning of internal layers in artificial neural networks from the perspective of neurons, shedding light on the black-box nature of the network's internal operations. There is research indicating a positive correlation between the PSI and performance in each layer of convolutional neural network models for image classification. In this study, we observed the internal operations of a convolutional neural network when adversarial examples were applied. The results of the experiments revealed a similar pattern of positive correlation for adversarial examples, which were modified to maintain 5% accuracy compared to applying benign data. Thus, while there may be differences in each adversarial attack, the observed PSI for adversarial examples demonstrated consistent positive correlations with benign data across layers.

• Journal of Korea Society of Industrial Information Systems
• /
• v.29 no.3
• /
• pp.23-40
• /
• 2024

Hardware attacks involve physical reverse engineering efforts to steal sensitive information, such as encryption keys and circuit designs. Encryption and obfuscation are representative countermeasures, but they are nullified if adversaries manage to find the key. To address this issue, we propose e-Cryptex, which utilizes a Physically Unclonable Function (PUF) as an anti-tampering shield. PUF acts as a random number generator and relies on unique physical variants that cannot be replicated or restored to enhance anti-tampering mechanisms. e-Cryptex uses PUF as a shield to protect the system's structure and generate the key. Tampering with the shield will result in the destruction of the key. This paper demonstrates that e-Cryptex meets PUF security requirements and is effective in detecting of tampering attempts that pierce or completely destroy the shield. Each board consistently generates the same key under normal conditions, while also showing key uniqueness across different boards.

• Journal of Korean Society of Steel Construction
• /
• v.16 no.6 s.73
• /
• pp.737-745
• /
• 2004

In general, most of the beam-to-column connections for steel structures are regarded as either rigid connections or pin connections. Recently, the concept of a semi-rigid connection was introduced for a correct analysis of steel structures. Several experimental and theoretical researches have been performed regarding the structural behaviors of frames and buildings with semi-rigid connections. The results are not well known, and structural frame/building has not been designed to introduce the concept of semi-rigid connections between a beam and column until this time. To resolve this, this research depends on design specifications prepared by other advanced countries for the design of buildings with semi-rigid connections. Such a specification, however, should incorporate domestic characteristics of steel material properties and load conditions. This paper deals with structural capacities and deformable behaviors for a split-T tensile connection with F10T high-strength bolts to investigate the structural characteristics of semi-rigid frames. The experimental parameters include the thickness of T-flanges, painted or not, preloaded or not, and load pushover pattern. A total of 20 specimens were fabricated and tested with a 300-ton UTM. The structural capacities and behavior for split-T tensile connections were evaluated on each research parameter.

• Journal of the Korean Magnetics Society
• /
• v.12 no.6
• /
• pp.218-223
• /
• 2002

We investigated the magnetic properties of High Flux-type $Ni_{x}Fe_{100-x}$(x=40∼50, wt.％) permalloy powders and dust cores. The powder was prepared by conventional gas atomization in mass production scale. At the composition of $Ni_{x}Fe_{55}$, saturation magnetization was maximum. In case of lower Ni content than X=45, the $M_{s}$, decreased largely with the decrease in Ni content, which is due to the invar effect. The permeability of compressed powder cores increased with the decrease in Ni content, which was considered to be due to the decrease in the magnetostriction. In addition, the dust core with Ni=45％ showed the lowest core loss because of the increase in electrical resistivity leading to the low eddy current loss. From the better frequency dependence of permeability, larger Q value and superior DC bias characteristics of Ni=45％ than those of Ni=50％ core, it was confirmed that the 45％Ni-55％Fe powder alloy was better material for the dust core than commercial High Flux core materials.

• Cartoon and Animation Studies
• /
• s.33
• /
• pp.55-78
• /
• 2013

The issues on the representation of digital images, the uncanny is not considered aesthetically but must be overcome. It seems that the future of visual media using digital images, depends on the development and advancement of the digital technology. However, the Brothers Quay regarded as unique and experimental animators, focus their attention towards the past rather than the future. They trace back to the origin of the media, reproducing, manipulating, and representing gestures in oder to restore the shock and thrill delivered in film and animation. The Brothers Quay explore the uncanny of gestures through the mechanism of medium. They work in a wide range of genres, from animation, live action to dance film. The creaking spasms, which create the feeling that the movement does not belong to this world, and the strange gestures, reveals the flicker of the still projection and make audiences recognize gestures in the true meaning of the word. While they animate live actors and dead objects like automata, the hidden desires are revealed over screen. In this paper, focusing on the works of Brothers Quay, I tried to explore the uncanny of gesture based on Freud's and Jentsch' theory of uncanny.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.21 no.5
• /
• pp.134-140
• /
• 2017

In this study, the flexural behavior of high strength concrete members with different curing condition of 90 MPa of compressive strength was investigated. Experimental parameters included normal and low temperature curing conditions, tensile steel amount and concrete compressive strength. 8 beam members were fabricated and flexural tests were carried out. Crack spacing, load-deflection relation, load-strain relation and ductility index were determined. Experimental results show that as the amount of rebar increases, the number of cracks increases and the crack spacing decreases. The higher the concrete strength, the smaller the number of cracks, but the effect is significantly smaller than the amount of rebar. As a result of comparison with the proposed average crack spacing in the design criteria, the experimental results are slightly larger than the results of the proposed formula, but the proposed formula does not reflect the concrete strength and curing conditions. The ductility index of normal temperature cured members was 3.36~6.74 and the ductility index of low temperature cured members was 1.51~2.82. The behavior of low temperature cured members was found to be lower than that of normal temperature cured members. As a result of comparing the ductility index with the existing studies similar to the experimental members, the ductility index of the high strength concrete member was larger than the ductility index of the ordinary strength concrete of the previous study. Further research is needed to understand more specific results.

• Journal of the Korea Concrete Institute
• /
• v.23 no.5
• /
• pp.657-665
• /
• 2011

LB-DECK, a precast concrete panel type, is a permanent concrete deck form used as a formwork for cast-in-place concrete pouring at bridge construction site. LB-DECK consists of 60 mm thick concrete slab and 125 mm height Lattice-girders partly embedded in the concrete slab. These decks have been applied to the bridges, which girder spacings are short enough to resist longitudinal cracking caused by construction loads. This paper presents experimental research work conducted to evaluate the cracking load of LB-DECKs designed for long span bridge decks. Twenty four non-composite beams and four composite beams are fabricated considering three design variables of thickness of concrete slab, height of lattice-girder, and diameter of top-bar. Static loads controlled by displacements are applied to test beams to obtain cracking and ultimate loads. Vertical displacements at the center of beams, strains of top-bar, crack propagation in concrete slab, and final failure modes are carefully monitored. The obtained cracking loads are compared to the analytical results obtained by elastic analyses. Long-term analyses using age-adjusted effective modulus method (AEMM) are also conducted to investigate the effects of concrete shrinkage on the cracking loads. Based on the test results, the tensile strength and the design details of LB-DECKs are discussed to prevent longitudinal cracking of long span bridge decks.

• Journal of the Korean Geotechnical Society
• /
• v.21 no.3
• /
• pp.19-25
• /
• 2005

A tunnel that uses the RMR method or the Q-system is called a 'modem tunnel' because the New Austrian Tunneling Method (NATM) is not employed, even though shotcrete and rock bolts are used as support. It is known that the modem tunnel, which is supported by shotcrete, is basically different from the conventional tunnel, which is supported by steel ribs. In order to preserve the load-carrying capacity of the rock mass, loosening and excessive rock deformations must be minimized. Although it is known that this can be achieved by applying shotcrete in the case of the modem tunnel, this has not been clearly demonstrated. In order to inspect the distinctions between the conventional tunnel and the modern tunnel, their support characteristics and the rock loads of the rock mass classifications are compared. Terzaghi's rock load classification was used as the conventional tunnel's representative rock mass classification. The RMR method and the Q-system were adopted as the modem tunnel's representative rock mass classification. The study's results show that the load-carrying capacity of shotcrete, when used as the main support in the modern tunnel, is greater than the load-capacity of the steel ribs used in the conventional tunnel. Because it has been verified that the rock loads of their rock mass classifications are not different, then, according to the rock mass classifications, the load-carrying capacity of the rock mass of the modern tunnel, which uses shotcrete, is not greater than that of the conventional tunnel.

• Journal of the Korean Geotechnical Society
• /
• v.15 no.6
• /
• pp.285-296
• /
• 1999

Preloading method is used to prevent the settling of a foundation and to increase the strength of ground by consolidation settlement in advance. But, the embankment used in preloading method brings large deformation and sliding failure in the soft ground. Recently, reinforcement method is often used in embankment in order to prevent sliding failure. But, until now, the research on the stability analysis considering both the rate of strength increase of clay by embankment load and increase of resistance force by the geosynthetics in the embankment body is not found. In this study, the stability analysis program(REAP) for embankment including these two points is developed. By this program(REAP), the stability analysis can be done about during the gradual increase of embankment and the stability counterplan can be established when the safety factor is lower than allowable safety factor of design. After calculating the position of sliding failure surface, the force of geosynthetics which is selected by either the effective tensile strength or tensile force caused by the displacement of soil mass in this position is applied to stability analysis. And the increase of resisting moment can be calculated by this force. Also, the construction period can be estimated and the time for the appropriate counterplan can be decided in order to maintain the stability of embankment. And then, safe and economical embankment design can be performed.

• Journal of The Korean Association For Science Education
• /
• v.33 no.1
• /
• pp.1-16
• /
• 2013

In this research, we tried to discover the characteristics of gifted students by analyzing their experience in showing task commitment. In order to do this, we asked science gifted students to specifically describe their experiences while deeply experimenting on a scientific cause or theory. From their responses, we inductively explored the characteristics of science gifted students by extracting and analyzing the characteristics that show task commitment. Consequentially, the characteristics of the gifted students are divided into nine categories, which are confidence, setting a challenging goal, challenging approach for solving problems, sense of potential control, loss of self-consciousness, time distortion, submission to difficult task, initiative, and endurance, all of which appear repeatedly among the gifted students. With consensus among three experts who have experience in research on gifted education, these nine characteristics can be categorized into 3 characteristics; challenge, flow, and willingness. The three characteristics such as challenge, flow, and willingness well represent a definition of task commitment. These characteristics can explain the level of task commitment exhibited by science gifted students. It is possible to develop the tool and framework for judging the task commitment of gifted students on the basis of their characteristics.